ERASMUS UNIVERSITY ROTTERDAM Reprint Prohibited Erasmus School of Economics
Master Thesis
Do Local Environments Shape the Fortune of Firms, or do Local Firms Shape a Fortunate Environment?
An evolutionary study of the geographical concentration of the U.S. automotive industry around Detroit.
Rick de Bruin
334452
Supervisor: Martijn J. Burger
Rotterdam, 16-11-2010
Content
1. Introduction
2. History of the industry
2.1The industrial climate of Detroit before 1890
2.2Geographic pattern of the U.S. automobile industry
2.3Market structure of the U.S. automobile industry
2.4Technological progress and the influence big Three
3.Related Literature
3.1Evolutionary economics
3.2Spinoff dynamics and agglomeration economies
3.3Windows of Locational Opportunity
3.4Related Industries
3.5The WLO concept and the U.S. automotive industry
3.6Hypotheses
4.Study Design and Data Description
4.1Data
4.1.1Sources of Data
4.1.2Checking Assumptions
4.2Data Description
4.2.1Automotive Industry Employment
4.2.2Firms
4.2.3Innovations
4.2.4Relatedness
4.2.5Control Variables
4.3Methodology
5.Empirical Findings
6.Conclusions
7.Limitations and Directions for further research.
8.References
9.Appendices
9.1Additional tables and figures.
9.2Data Description
9.3Checking Assumptions.
List of Tables
Table 1 Transmission type as a share of sales
Table 2 Theories of technological change
Table 3 Location Quotients for the North East and Mid West
Table 4 Automobile firms over states
Table 5 Number of Innovations over states
Table 6 Regression Results for analysis 1
Table 7 Regression Results for analysis 2, sensitivity analysis
Table 8 Regression Results for analysis 3, interaction effects
List of Figures
Figure 1 U.S. geographical distribution of automobile-assembly plants, 1900
Figure 2 U.S. geographical distribution of automobile-assembly plants, 1914
Figure 3 Number of entrants, exits and firms in the U.S. automobile industry
Figure 4 Annual sales of passenger cars
Figure 5 Percentage of firms located in Detroit
Figure 6 The window of locational opportunity concept
1.Introduction
Detroit is often used as a synonym for the American automotive industry and referred to as Motor City. In the period from 1900 through 1930 the city of Detroit experienced an exceptionally high physical growth for a large city and underwent an unprecedented pace of social and technological change. The flourishing of the Detroit area was mainly due to the large concentration of the automobile industry.
Initially the automobile industry was not confined to a single location, but rather was evolving in a number of widely scattered places, with the northeast as the centre of activity (Klepper, 2007). Most interestingly is the fact that in 1895, forty firms were producing motor vehicles for commercial purpose, however none of these firms was located in Detroit (Boas, 1961). The population of Detroit grew enormously, from 285,704 in 1900 to 1,568,662 in 1930. According to Klepper (2007) this unparalleled population growth of Detroit can be explained by the concentration of the automobile industry around Detroit which shifted from the Atlantic coastal region towards Detroit and Chicago supported by cities surrounding the Great Lakes such as Indianapolis, Cleveland and Milwaukee.
Rubenstein (2002) confirms this shift in automotive industrial activity when he found that best selling cars were made in southeastern Michigan. By 1914 a total of 50 assembly plants established Detroit as the capital of the automobile industry accounting for 80% of the total industry output with respect to leading makes and giving home to the three famous Detroit-based firms, General Motors, Ford and Chrysler also known as the big Three (Hurley, 1959). When the number of firms reached its top in 1909 at 274 firms entry stayed stable for a few years but started to decline in 1910 (Klepper and Simons, 1997, Klepper, 2007). The industry became more competitive as larger firms benefited from their economies of scale and mass production drove down the price of a motor vehicle (Klepper, 2007). Detroit consolidated its supreme position as a leader in the U.S. manufacturing car industry by absorbing many of the marginal producers inducing an industry shakeout which changed the market structure into a tight oligopoly dominated by the “Big Three”.
The geographical pattern which characterizes different industries gained particular interest of many schools in economics, geography, history and management (Boschma et al., 2002). Besides the automobile industry there are several industries, as the US tire industry around Akron, Ohio and the famous example of Silicon Valley which show similar patterns of concentration (Klepper, 1997). These agglomerations may provide insight in making a distinction between favourable location advantages and creative ability of firms in terms of generating and attracting their own favourable production environment. Getting insight in these forces could resolve a part of mystery in agglomeration economies and what should be the driving forces of endogenous growth and real business cycles (Ellison and Glaeser, 1999).According to geographic economics industries tend to concentrate and agglomerate in a region. However few other industries, then the U.S. automobile industry, are so closely identified with one single geographic location as in the case of Detroit. Many researchers have tried to identify the driving forces of the concentration and agglomeration of the U.S. automobile industry. Unfortunately, there is no consensus on why Detroit became the capital of the automobile industry. The current literature can roughly be divided in two bodies of research were one of them posits as central the importance of static location factors for the rise and agglomeration of new industries. Ellison and Glaeser (1999) found that industries’ locations are influenced by a wide range of natural advantages. In the case of the Detroit area researchers refer to Detroit’s low-cost access by water to raw materials. However Hurley (1959) argued that the automotive industry wasdependent on any constant location specific factors such as land, climate or raw materials. Though he emphasized the importance of the flat land and the presence of the Great Lakes. According to Hurley these natural advantages resulted in a high concentration of vital industries which were extremely important in the early phase of the automotive industry. The indirect effects of natural advantages led to a large supply of basic suppliers and pools of semi-skilled and skilled workers.
Another explanation is given by Rae (1965), who attributes the supremacy of Detroit in the U.S. automobile industry to the fact that the city possessed “a unique group of individuals with both business and technical ability who became interested in the possibilities of the motor vehicle”. Rae (1965) referred to pioneers of the U.S. automobile industry such as Henry Ford, Ransom E. Olds, Elwood Haynes and Charles Duryea who lived in Michigan and by some extend became active in the region of Detroit.
Once these pioneers were located in the Detroit area it is plausible that agglomeration economies alone could have accounted for the subsequent growth of the industry in the Detroit Area. These agglomeration economies led to better infrastructure, the emergence of specialized suppliers, a large pool of semi-skilled and skilled workers and the establishment of supportive institutions (Boschma, 2007). However if these driving forces of agglomeration were significant, and these positive externalities were geographically bounded, firms of all types in Detroit should have benefitted from these supportive conditions and performed comparably with the automotive industry[1].
The driving forces behind the supremacy of Detroit as the capital of the U.S. automobile industry goes beyond the scope of static agglomeration economies. This means that the emergence of new industries is not a process in which rational economic actors choose their location based on the existing locational structure in order to minimize production costs. It would be misleading to interpret the spatial formation of a new industry based on how well the new requirements best correspond to specific locational factors. Therefore the main focus of this article will be on the second body of research which takes a critical stand towards the interpretation of spatial formation as an allocation process where the industry develops in places dependent on the existing local structures. It views the spatial formation of a new industry as a dynamic process of local development.
With this article I hope to provide new insights into the nature of geographic concentration of new industries. The main goal is to further enhance our understanding on whether new industries reflect continuous, evolutionary rather than discontinuous changes in spatial economies.
By extending the windows of locational opportunity concept by Boschma (Boschma, 1996) and the spinoff theory by Klepper (2007), this article provides new insights in to which extend the high concentration of the U.S. automobile industry in Detroit was due to the initial favorable local environment of Detroit or the ability of the emerging automotive industry to create its own supportive local environment. The theoretical concepts by Boschma (1996, 2002) and Klepper (2002a, 2007) regarding the varying ability of regions to grasp new technological opportunities with respect to technology and labor will be tested by, respectively hypotheses on the geographic distribution of innovations and professions in the automobile industry. The study tries to identify different stages of development of the U.S. automotive industry by making a distinction between the evolutionary concept of innovation and the discontinuous revolutionary concept of innovation. The evolutionary concept of innovation is composed by an accumulation of technological knowledge, skills and experience while the discontinuous revolutionary concept of innovation reflects a dramatic break with the past with regard to techno-industrial development process.
The structure of this article is as following; Section 2 will start with a historical overview of the U.S. automobile industry followed by section 3 which gives a theoretical framework with theories on dynamic and evolutionary economics with respect to industrial location. The focus will lie on the window of locational opportunity concept whereby the role of local labor markets and innovations is made explicit. Also limitations of these theories, the research question as well as the hypotheses will be presented. In section 4 the data, methodology, and the models which have been applied will be presented. Also a proper explanation of the used variables will be given. Section 5 presents some results of the analyses on the geographic distribution of innovative activity and the presence of related industry employment prior to the emergence of the automotive industry.. Section 6 will draw some conclusions and elaborates on the scientific and social relevance of the main findings. Section 7 will end with limitations and suggestions for further research.
2.History of the industry
According to Ellison and Glaeser (1997) industries are typically agglomerated geographically. However the exceptional high concentration of the automobile industry in Detroit, Michigan goes beyond the explanation of agglomeration economies. While the entire world considers Detroit as the capital of the U.S. automobile industry this was not necessarily the case throughout the history of the automotive industry. In this section we analyze the history and development of the U.S. automotive industry and its geographical concentration around Detroit. Subsection 2.1 gives an overview of the industry structure of Detroit before the start of the automobile industry. In subsection 2.2 the geographical pattern of the U.S. automobile industry, changing over time, is analyzed. 2.3 analyses the evolution of the market structure volume and size of producers. The largest breakthroughs in terms of product and process innovations that accelerated the car industry, as well as the influence of America’s biggest car manufacturers is handled in subsection 2.4. Subsection 2.5 will draw some conclusions.
2.1The industrial climate of Detroit before 1890
When answering the question were and why new industries should emerge there is much debate about whether the spatial manifestation of innovations can be determined by specific circumstances or are the outcome of chance events (Boschma, 1996). This is because new techno-industrial knowledge can hardly draw on local conditions to support its economic exploitation as the new requirements of such a process are not given but come gradually into being as the development of the industry proceeds. With respect to the automotive industry in the United States, one might wonder if the extreme high concentration around Detroit can be explained by its former industrial structure. The following section describes the industrial climate of Detroit before the start of the automotive industry at the end of the 20th century. According to Denison (1956) the strength of the industry around Detroit finds its foundation in geology. The withdrawal of gigantic ice sheets thousands of years ago shaped a landscape providing favorable conditions. The presence of the Great Lakes, gave Detroit direct access to economic assets like iron, copper, salt and hardwood. Naturally, the abundance of these raw materials, combined with different skills resulted in a diversified industry structure producing alarge scope of different goods. The Detroit area became famous for its steel railroad rails, wagons, carriages, all sorts of woodenwares but especially because of its stoves and smallmarine engines powered by illuminating gas and naphtha (Denison, 1956). Michigan became a leading center for the production of gasoline engines for agricultural and marine uses. Detroit’s shipyards were early converts to gasoline engines because of their dissatisfaction with the heating capacity of coal (Rubenstein, 2002). Denison (1956) also devotes the population of Detroit as a great asset in the development of the automotive industry. A mixed composition of people resulted in a diversified set of labor, skills and talent which made it possible to expand in different industries. As mentioned Detroit became a pioneer in making steel, railroad cars, industrial chemicals, carriages and marine engines. However, according to Denison (1956) and Hurley (1959), Detroit did not have a single dominating industry and instead of committing to a single resource the city created the perfect framework to grasp future opportunities. Denison (1956) and Hurley (1959) refer to fact that Detroit became the capital of the U.S. automobile industry with the highest concentration of producing activity. However this was the result of an evolutionary process rather than a single grasped opportunity.That is, the invention and economic exploitation of the automobile did not fall from the sky. Furthermore Detroit did not produce cars for commercial purpose for only 7 years after commercial production started in Connecticut and New York (Smith, 1968). The next section will provide an insight in this evolutionary process and provides an overview in the changes over time of the geographic pattern of the U.S. automobile industry.
2.2Geographic pattern of the U.S. automobile industry
Based on studies of Boas (1961), Smith (1968) and Klepper (2007), the following section will give an overview of the geographic pattern of the U.S. automobile industry during the period 1895-1958. The first company that was organized to produce motor vehicles for commercial purpose was the Duryea Motor Wagon Company of Springfield, Massachusetts. After the leading start of Duryea in 1885, commercial production began and activity was scattered across the northeastern United States. Of the 40 plants operating in 1895 the only significant concentration was the grouping of 7 plants in the region around New York City. Interestingly enough there was no activity in the automobile industry in the city of Detroit. In the last five years of the 19th century the automobile industry grew rapidly and reached a total of 327 plants. By 1900 the number of plants in the Detroit area was negligible and the highest concentration, more than half of all automobile-assembly plants were located in the Atlantic coastal region as can be seen in figure (1). In 1905 a major shift in the locational pattern of the U.S. automobile industry started. Despite the fact that Boston, New York, northern New Jersey, Philadelphia and the upstate New York city’s continued to be the main centers of the industry, Chicago, Cleveland, Detroit, Milwaukee and St. Louis showed increasing automotive manufacturing activity. This trend towards the Midwest continued and between 1908 and 1911 there was a strong decline in assembly plants in the Atlantic Seaboard area and the concentration of assembly plants in San Francisco had disappeared. By 1914 the total number of car manufacturing firms reached its peak with 469 active assembly plants. More importantly Detroit became the Capital of the U.S. automobile industry, giving home to 50 assembly plants, making 1914 one of the most important years in the early phase of the automobile industry. The centre of the automobile industry was shifted from the Atlantic coastal region towards Detroit and Chicago supported by cities surrounding the Great Lakes like Indianapolis, Cleveland and Milwaukee as illustrated in figure 2. Despite the fact that the number of operating companies in 1919 was less than half of total assembly plants in 1914. After a small upturn in 1921 the industry started to decline and became more competitive. The larger firm started consolidating by absorbing many of the small manufacturers.
By 1926 the industry has been reduced to a total of 182 assembly plants the southern of Michigan had become the supreme center of the U.S. automobile industry. Although the economic crisisof the early 1930’s caused a large reduction in the total number of firms,the geographical distribution of automobile firms hardly changed. The main question that arises by looking at the development of the geographic distribution of U.S. automobile manufacturing plants is: Why did Detroit became the capital of the U.S. car industry? To illustrate the dynamic development the U.S. automobile industry underwent the following section will give a short overview of the changes in the automobile market structure during the period between 1895 and 1960.
Figure1U.S. geographical distribution of automobile-assembly plants, 1900. (Boas, 1961)
(Each dot representing one operating plant)
Figure 2 U.S. geographical distribution of automobile-assembly plants, 1914. (Boas, 1961)
(Each dot representing one operating plant)